Handbook mirroring section

[Warren Block]

On Mon, 11 Jun 2012, Hiroki Sato wrote:

> wb> I tried a whole-disk GPT mirror for a short time, and had no problems
> wb> other than gptboot complaining about the corrupted secondary header.
> wb> It was not an exhaustive test, though.
>
> Can you try 9-STABLE after Apr 26?  The warning should not appear.

In a FreeBSD-9-stable snapshot from June 12 (from your allbsd.org site, 
thanks!):

In FreeBSD, the loader did not complain with GPT inside a whole-disk 
gmirror.  There were warnings from GEOM when GEOM-level stuff happened 
to that drive.  Clear the gmirror metadata and see "GEOM: ada1: the 
secondary GPT header is not in the last LBA".  If this is really 
standards-compliant, GEOM should not complain about it so much.

GParted, the Linux graphical partition editor, saw the disk as GPT.
When parted, the command-line version, was told to print the disk data, 
it said that the secondary partition table was not at the end of the 
drive and offered to "fix" it; then it offered to "fix" the partition 
wasting that one block at the end of the drive.  Then it finally 
displayed the information.

On FreeBSD, I don't think there would be a problem, but there are so 
many different possibilities it's hard to think of a complete test.